Replace some trivial WebKitSystemInterface wrappers with direct calls to SPI.

Replace some trivial WebKitSystemInterface wrappers with direct calls to SPI.

BUG=101156
R=thakis@chromium.org

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=149556

[Nico, 2013-05-01 01:02:18]

lgtm with comments

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
File Source/core/platform/graphics/mac/FontCacheMac.mm (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
Source/core/platform/graphics/mac/FontCacheMac.mm:68: // Note: This may break on
future OS releases, per https://bugs.webkit.org/show_bug.cgi?id=102405#c8.
Is there a rdar for getting a public api for this? If not, please file one (and
then add a link to it here)

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
File Source/core/platform/graphics/mac/GlyphPageTreeNodeMac.cpp (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
Source/core/platform/graphics/mac/GlyphPageTreeNodeMac.cpp:65:
CGFontGetGlyphsForUnichars(fontData->platformData().cgFont(), buffer,
glyphs.data(), bufferLength);
Same question.

https://codereview.chromium.org/14574002/diff/6001/Source/core/rendering/Rend...
File Source/core/rendering/RenderThemeChromiumMac.mm (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/rendering/Rend...
Source/core/rendering/RenderThemeChromiumMac.mm:747: _NSDrawCarbonThemeBezel(r,
isEnabled(o) && !isReadOnlyControl(o), YES);
This is ok since this seems to be unnecessary on 10.8+

[eseidel, 2013-05-01 02:11:03]

One thing we lose with this is the ability to know what SPIs we actually use. 
I'm not sure how useful that information is.  But an alternative to this would
be to just keep a single header (we could call it MacOSXSPIs.h or similar?)
which gave us all these forward declarations?

I don't feel particularly strongly one way or the other.

[eseidel, 2013-05-01 03:19:39]

I think our ability to influence Apple (to add/keep SPIs) is basically nil. 
WebKit had some influence, with an internal team, but I think as you say, we
should just use the SPIs if they're there and deal when they're not. :)  Having
a central list won't help us in that.

[Robert Sesek, 2013-05-01 17:05:39]

I can file the radars if you think it's really necessary, but my guess is we
won't get a public way to do this. We'll need to keep these SPIs around for
considerable time, anyways.

Regarding a centralized list of SPIs: this is not how we've done things in
Chromium, and as you point out Eric, it doesn't buy us much. There are a few
more complex methods (take a privately defined struct) that are used in multiple
files, and I will put those in a shared header. But for just simple functions,
declaring in the file that it's used I think is the way to go. I'm leaving
comments with "SPI," though, so it's at least greppable.

[eseidel, 2013-05-01 18:31:07]

I'm not sure it matters much. Any we inherited from WebKit should already have
radars for them.  But yes, the "right" thing is to file radars.  Despite radar
being a black hole...

[Robert Sesek, 2013-05-01 23:58:43]

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
File Source/core/platform/graphics/mac/FontCacheMac.mm (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
Source/core/platform/graphics/mac/FontCacheMac.mm:68: // Note: This may break on
future OS releases, per https://bugs.webkit.org/show_bug.cgi?id=102405#c8.
On 2013/05/01 01:02:18, Nico wrote:
> Is there a rdar for getting a public api for this? If not, please file one
(and
> then add a link to it here)

Done.

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
File Source/core/platform/graphics/mac/GlyphPageTreeNodeMac.cpp (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/platform/graph...
Source/core/platform/graphics/mac/GlyphPageTreeNodeMac.cpp:65:
CGFontGetGlyphsForUnichars(fontData->platformData().cgFont(), buffer,
glyphs.data(), bufferLength);
On 2013/05/01 01:02:18, Nico wrote:
> Same question.

Done.

https://codereview.chromium.org/14574002/diff/6001/Source/core/rendering/Rend...
File Source/core/rendering/RenderThemeChromiumMac.mm (right):

https://codereview.chromium.org/14574002/diff/6001/Source/core/rendering/Rend...
Source/core/rendering/RenderThemeChromiumMac.mm:747: _NSDrawCarbonThemeBezel(r,
isEnabled(o) && !isReadOnlyControl(o), YES);
On 2013/05/01 01:02:18, Nico wrote:
> This is ok since this seems to be unnecessary on 10.8+

Correct.

https://codereview.chromium.org/14574002/diff/6001/Source/core/rendering/Rend...
Source/core/rendering/RenderThemeChromiumMac.mm:782:
_NSDrawCarbonThemeListBox(r, isEnabled(o) && !isReadOnlyControl(o), YES, YES);
Filed.

[Robert Sesek, 2013-05-02 00:03:45]

Committed patchset #3 manually as r149556 (presubmit successful).